I have virtually no experience with OCR in XSane or other Linux scanning software, because until recently, every document I've ever scanned in a Linux program contained no text. I have an old (twelve years old) typed letter, and decided to scan it and use OCR.

The results were very disappointing. It's 90% gibberish! Obviously I can't settle for it. Do you get results like this when you try to OCR a document in Xsane?

Or is it probably not the software but the scanner? My scanner's used, I acknowledge.

I hadn't thought of the obvious--increase the resolution. It had been only 75. That helps very much, but the results are still disappointing. I pushed it all the way up to 1200 (maximum on this scanner), but I;m still reading plenty of errors.

Since I don't use OCR much, I have no idea if it's reasonable to expect 100 percent accuracy or if the technology is a work in progress.

Let's let XSane tell the story,

http://xsane.org/xsane-introduction.html

It doesn't look like Xsane has been upgraded in many years: http://xsane.org/index.html

The SANE Library,

http://www.sane-project.org/

2 members found this post helpful.